A method for allocating power between satellite links in heterogeneous scenarios

Abstract

The present invention provides an inter-satellite link power allocation method for heterogeneous scenarios, including: updating the inter-satellite link gain parameters between the current medium and low orbit satellites; updating the remaining visible time convergence coefficient and the remaining visible time penalty coefficient respectively according to the orbital period and normalized throughput of the low orbit satellite, and then obtaining the remaining visible time cost penalty function; obtaining the power output result of each low orbit satellite after the game based on the overall utility game; using the power output result to calculate the utility value and clustering threshold of each low orbit satellite in the cluster, and performing clustering processing on the low orbit satellite whose utility value is lower than the clustering threshold; entering the next time slot, updating the remaining visible time of all low orbit satellites, and initializing the remaining visible time of all low orbit satellites newly added to the cluster. The inter-satellite link power allocation method for heterogeneous scenarios of the present invention optimizes the overall performance of the medium and low orbit satellite network by designing a reasonable utility reward and punishment function and a clustering threshold scheme.

Description

Technical Field

[0001] The present invention belongs to the field of satellite communication architecture design, and in particular relates to an inter-satellite link power allocation method for heterogeneous scenarios. Background Art

[0002] Non-terrestrial networks are an extension of terrestrial networks, used to supplement remote areas that cannot be served or are difficult to be effectively served by terrestrial networks. However, due to the special geographical nature of non-terrestrial networks, they have high technical requirements. At the same time, satellites in non-terrestrial networks are expensive, which makes the number of non-terrestrial network satellites far lower than the number of terrestrial network equipment. On the other hand, my country’s satellite resources are insufficient, and most medium and low-orbit satellite orbital resources have been divided up by other countries. Therefore, how to utilize limited medium and low-orbit satellite resources and make full use of the high dynamic characteristics of medium and low-orbit satellite orbits to build an efficient medium and low-orbit constellation Internet network to provide users with low-latency and seamless service experience to the greatest extent is an issue that needs to be solved urgently.

[0003] When low-orbit satellites and medium-orbit satellites interact with each other in the same network, the transmission power of each low-orbit satellite will interfere with other low-orbit satellites in the network. However, any low-orbit satellite will increase its signal-to-noise ratio by increasing its own transmission power, which will cause serious interference to other low-orbit satellite members in the same network, significantly reducing the overall network performance and dramatically increasing satellite power consumption, resulting in a waste of resources. At the same time, due to the high dynamic operation characteristics of low-orbit satellites and the complex topological structure of low-orbit satellite networks, low-orbit satellites in the same network will frequently enter each other's blind spots. If the low-orbit satellites in the blind spots cannot be managed efficiently and timely, satellite resources will also be wasted.

[0004] At present, the common on-board power allocation methods are mostly borrowed from ground networks, mainly including greedy power allocation schemes and water injection power allocation schemes, but they mostly consider the performance of each user, while ignoring the interference between users, and cannot ensure the optimal overall performance of the network. At the same time, in view of the time-varying characteristics of low-orbit satellites, there are two common strategies for when low-orbit satellites leave the network:

[0005] 1) Store the specific orbital position and period information of low-orbit satellites. When a low-orbit satellite passes a certain spatial point, it will actively trigger the low-orbit satellite to leave the network. If there are multiple complex time-varying low-orbit satellites in the network, these low-orbit satellites interfere with each other. Due to the influence of interference and other factors, it cannot be guaranteed that these low-orbit satellites can benefit the overall performance of the network when they leave the network at a fixed point. Therefore, this strategy is only applicable to the situation where the low-orbit satellite members in the network are determined and single.

[0006] 2) Conduct pilot channel measurement in real time. When the channel quality is lower than a certain threshold, the low-orbit satellite is triggered to leave the network. However, continuous pilot monitoring by the low-orbit satellite will greatly increase the satellite's own power consumption, which is not applicable to the situation where the low-orbit satellite has limited battery.

[0007] Therefore, a reasonable power allocation solution needs to be established for complex heterogeneous scenarios. This solution can balance the considerations of each user in the network and improve the overall network performance for different scenarios. Summary of the invention

[0008] The purpose of the present invention is to provide an inter-satellite link power allocation method for heterogeneous scenarios, so as to reasonably allocate power resources to low-orbit satellites and optimize the overall performance of medium and low-orbit satellite networks.

[0009] In order to achieve the above object, the present invention provides an inter-satellite link power allocation method for heterogeneous scenarios, comprising:

[0010] S1: Set the dynamic scene of low-orbit satellites relative to medium-orbit satellites, initialize the number of low-orbit satellites K in the current cluster; set the current time slot k=1, set the time slot interval Δt; set the initial remaining visible time of N low-orbit satellites in the cluster to the medium-orbit satellite; set the number of power levels M and the transmission power vector of low-orbit satellites;

[0011] S2: If K≠0, update the inter-satellite link gain parameter h(t k )=[h 1 (t k ) 2 (t k )…h K (t k )]; According to the orbital period of the K low-orbit satellites in the cluster at the current time slot k, the normalized throughput C(t k ) respectively update the remaining visible time convergence coefficient k i1 and the remaining visible time penalty coefficient k i2 , and then get the remaining visible time penalty function Cost time (t k ); if K = 0, go to step S5;

[0012] S3: If K=1, proceed to step S4; otherwise, proceed to step S6;

[0013] S4: Output the best utility of the only low-orbit satellite in the current time slot k and its corresponding power value, and enter step S11;

[0014] S5: Set the best utility of the current time slot k to zero and proceed to step S11;

[0015] S6: Fix the transmission power p of the i-th low-orbit satellite in the current time slot k ti (t k ), traverse the transmission power of the other K-1 low-orbit satellites in the cluster except the i-th low-orbit satellite The overall utility of the ith low-orbit satellite and all low-orbit satellites except the ith low-orbit satellite in the current time slot k is calculated under different combinations of transmission power, and the maximum value is selected to form the initial overall maximum utility array u of the ith low-orbit satellite in the current time slot k ini_max_i , the transmission power of each of the K LEO satellites in the corresponding cluster is stored in the initial optimal transmission power response array p of the i-th LEO satellite ini_best_i_q_matrix , the corresponding transmission power of the i-th low-orbit satellite constitutes the initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i ;

[0016] S7: i = i + 1, and return to step S6 until all low-orbit satellites in the cluster are traversed, and the initial overall maximum utility set u is obtained. ini_max and the corresponding initial optimal transmit power response set, and continue to execute step S8;

[0017] S8: respectively calculating the minimum distance between the matrix of the initial best transmit power response array of the low-orbit satellite used as the reference satellite in the initial best transmit power response set and the initial best transmit power response array of any other low-orbit satellite, and selecting the transmit power combination corresponding to the minimum distance as the power output result after the game of each low-orbit satellite in the cluster;

[0018] S9: Calculate the utility value and cluster-leaving threshold of each low-orbit satellite in the cluster at the current time slot k using the power output results of each low-orbit satellite in the cluster after the game. If the utility value of any low-orbit satellite is lower than the cluster-leaving threshold, the low-orbit satellite will be de-clustered. At this time, K = Kn, where n refers to the number of satellites whose utility value is lower than the cluster-leaving threshold.

[0019] S10: The power output results of all low-orbit satellites in the cluster that have not left the cluster after the game are taken as their optimal transmission power, and the sum of the utility values ​​of all low-orbit satellites in the cluster at the current time slot k is calculated and taken as the optimal overall utility of the current time slot;

[0020] S11: let k=k+1; then, if K>0, update the remaining visible time of all low-orbit satellites in the cluster respectively, and then enter step S12: if K=0, directly enter step S12;

[0021] S12: Initialize the remaining visible time of all low-orbit satellites newly added to the cluster, and use K+l as the updated K, where l is the number of low-orbit satellites newly added to the cluster in the new current time slot;

[0022] S13: If all time slots have been traversed, the process ends, otherwise it returns to step S2.

[0023] In step S1 and step S12, the initial remaining visible time t is:

[0024] t=[t 10 t 20 …t N0 ], t i0 is the initial remaining visibility time of the i-th LEO satellite to the ME satellite in the cluster;

[0025] t i0 The value range of is:

[0026] 0≤t i0 ≤|t i2 -t i1 |,

[0027] Among them, t i1 ,t i2 is the time when the visibility of the i-th low-orbit satellite switches between two consecutive times, t i1 represents the time when the i-th low-orbit satellite enters the visible area of ​​the medium-orbit satellite, t i2 represents the time when the i-th low-orbit satellite leaves the visible area of ​​the medium-orbit satellite;

[0028] In step S1, the transmission power range of the low-orbit satellite is set [p min , p max ], set the transmission power step length Δp, and obtain the number of power levels M = (p max -p min ) / Δp, the transmission power vector of the low-orbit satellite is obtained as: [p min p min +Δp…p max -Δp p max ].

[0029] The utility value u of the i-th low-orbit satellite in the cluster at the current time slot k is i (t k )for:

[0030]

[0031] Among them, R throughput (t k ) is the profit function at the current time slot k, Cost energy (t k ) is the energy cost penalty function at the current time slot k, Cost time (t k ) is the remaining visible time cost penalty function.

[0032] The remaining visible time penalty function Cost of n low-orbit satellites in the cluster at the current time slot k time (t k )for:

[0033]

[0034] Among them, the coefficient k i1 is the remaining visible time convergence coefficient, i∈(1, 2, …, n), which determines the speed of convergence of the remaining visible time cost penalty function, k i2 Defined as the remaining visible time penalty coefficient, k is the ordinal number of the current time slot, and Δt is the time slot interval;

[0035] k i1 ∝1 / T i (k i1 >0), k i2 ∝1 / log(1+ξ i (t k ))(k i2 >0),

[0036]

[0037] Among them, T i is the orbital period of the i-th LEO satellite in the cluster, log(1+ξ i (t k )) is the throughput of n low-orbit satellites in the cluster at the current time slot k, ξ i (t k ) is the signal-to-noise ratio of the i-th LEO satellite in the cluster at the current time slot k;

[0038] The clustering threshold of the i-th low-orbit satellite in the cluster is:

[0039]

[0040] Among them, μ is the weight factor, and the value range of μ is 0.01-0.5.

[0041] Energy consumption penalty function Cost at the current time slot k energy (t k )for:

[0042] Cost energy (t k )=[ηE 1 (t k ) ηE 2 (t k )…ηE n (tk )],

[0043] Among them, η is the energy consumption penalty factor, E i (t k ) is the normalized energy consumption of the i-th low-orbit satellite in the current time slot k.

[0044] The profit function R at the current time slot k throughput (t k )for:

[0045] R throughput (t k )=[λC 1 (t k ) λC 2 (t k )…λC n (t k )],

[0046] Where λ is the throughput benefit factor, C i (t k ) is the normalized throughput C(t k );

[0047] The normalized throughput C(t k )for:

[0048] C(t k )=[C 1 (t k ) C 2 (t k )…C n (t k )] = β·[log(1+ξ 1 (t k )) log(1+ξ 2 (t k ))…log(1+ξ n (t k ))],

[0049] Among them, C i (t k ) is the normalized throughput of the i-th LEO satellite in the cluster at the current time slot k, β is the normalization factor, ξ i (t k ) is the signal-to-noise ratio of the i-th LEO satellite in the cluster at the current time slot k, ξ i (t k )=p ti (t k ) / σ 2 , p ti (tk ) is the transmission power of the i-th LEO satellite in the cluster at time slot k, σ 2 is the white noise power in the satellite signal transmission space.

[0050] The initial overall maximum utility array u of the i-th low-orbit satellite at the current time slot k is ini_max_i for:

[0051]

[0052] Among them, u ini_max_i_q It refers to the initial overall maximum utility value of the i-th low-orbit satellite under the q-th transmission power combination, 1≤q≤M K-1 , M K-1 represents the number of combinations of the transmission power of the other K-1 low-orbit satellites in the cluster, and M is the number of power levels;

[0053] The initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i for:

[0054]

[0055] Among them, p ini_best_i_q is the initial optimal transmission power of the i-th low-orbit satellite under the q-th transmission power combination.

[0056] The initial overall maximum utility value u of the i-th low-orbit satellite under the q-th transmission power combination ini_max_i_q and the corresponding initial optimal transmission power array p ini_best_i The calculation process of is as follows: take a set of transmission power combinations of K-1 low-orbit satellites in the cluster except the i-th low-orbit satellite as the q-th transmission power combination, traverse the different transmission powers of the i-th low-orbit satellite under the condition of the q-th transmission power combination, and calculate the utility value u of the i-th low-orbit satellite in the cluster at the current time slot k i (t k ) and the calculation formula for the overall utility of the ith low-orbit satellite and all low-orbit satellites except the ith low-orbit satellite in the current time slot k Calculate the overall utility of the i-th low-orbit satellite at different transmission powers The maximum value of the overall utility is selected as the initial overall maximum utility value u of the i-th low-orbit satellite under the condition of the n-th transmission power combination. ini_max_i_q At this time, the corresponding transmission power of the i-th low-orbit satellite is the initial optimal transmission power p of the i-th low-orbit satellite under the q-th transmission power combination. ini_best_i_q At the same time, the transmission power of each of the K low-orbit satellites in the cluster is stored in the initial optimal transmission power response array p of the i-th low-orbit satellite ini_best_i_q_matrixIn the example, the initial optimal transmission power p for the low-orbit satellite i under the qth transmission power combination is ini_best_i_n Supplement.

[0057] The step S8 comprises: taking the first low-orbit satellite as a reference satellite, selecting an initial optimal transmission power response array p corresponding to different transmission power combinations of the reference satellite and the i-th low-orbit satellite ini_best_i_q_matrix The initial optimal transmission power response array pair p with the smallest distance between ini_best_1_m_matrix 、p ini_best_i_n_matrix , and output the initial overall maximum utility values ​​corresponding to the transmission power combinations m and n of the initial optimal transmission power response array pair, respectively, as the power output results p of the reference satellite after the game ini_best_1 (m) and the power output result p after the game with the i-th low-orbit satellite ini_best_i (n); then, the same process is repeated under the condition of changing the size of i to calculate the power output results of other low-orbit satellites in the cluster and the reference satellite after the game.

[0058] The intersatellite link power allocation method for heterogeneous scenarios of the present invention combines the low-orbit satellite's own orbit and throughput to design the remaining visible time cost penalty function Cost time (t k ), a greater utility penalty is imposed on the LOR satellite with less remaining visible time. If the channel quality of the LOR satellite is relatively good, the utility penalty is reduced accordingly. If the LOR satellite is in the blind spot of the medium-orbit satellite, a particularly large utility penalty is imposed, thereby solving the problem of power resource allocation for LOR satellites in a complex heterogeneous state.

[0059] In addition, the intersatellite link power allocation method for heterogeneous scenarios of the present invention designs the clustering threshold of the utility of low-orbit satellites on the basis of the game of overall utility. The size of the clustering threshold depends on the weight of the current satellite throughput benefit and the energy consumption penalty, so that the low-orbit satellite participants with utility lower than the threshold are placed in the clustering state to ensure the overall utility of the low-orbit satellite network in the cluster. If the throughput weight is larger, the clustering strategy is more conservative, and the low-orbit satellite is less likely to be clustered. On the contrary, if the energy consumption weight is larger, the clustering strategy is more radical, and the low-orbit satellite is more likely to be clustered.

[0060] In summary, the inter-satellite link power allocation method for heterogeneous scenarios of the present invention, by designing a reasonable utility reward and punishment function (mainly including designing a satellite remaining visible time cost penalty function) and an effective utility out-of-cluster threshold, enables the medium and low-orbit satellite networking network to maximize the performance and online rate of the low-orbit satellites in the cluster, while avoiding the abnormal situation that the low-orbit satellites in the cluster are in a communication blind spot but still remain in the cluster, thereby optimizing the overall performance of the medium and low-orbit satellite network. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is the inter-satellite network topology diagram for the joint networking of multiple medium and low-orbit satellites.

[0062] Figure 2 Diagram of the inter-satellite operation model for the same medium and low-orbit satellite network.

[0063] Figure 3 The present invention is an algorithm flow chart of the inter-satellite link power allocation method for heterogeneous scenarios. DETAILED DESCRIPTION

[0064] The present invention is described in detail below with reference to the accompanying drawings and specific examples.

[0065] The present invention proposes an inter-satellite link power allocation method for heterogeneous scenarios. The present invention is applicable to medium and low-orbit satellite communications in inter-satellite networks of various different dynamic scenarios, and the inter-satellite network is composed of a plurality of medium and low-orbit satellite networking networks.

[0066] like Figure 1 The figure shows the topological structure of the intersatellite network formed by multiple medium and low orbit satellites. Figure 1 As shown, each medium-orbit satellite network has a medium-orbit satellite as a manager, which is responsible for managing several low-orbit satellites as participants in the network and can communicate with any low-orbit satellite as a participant in the network. The medium-orbit satellite cannot communicate with low-orbit satellites outside the network. The same medium-orbit satellite network is referred to as a cluster below. The low-orbit satellites in the cluster can only communicate with the medium-orbit satellites in the cluster, but cannot communicate directly with the medium-orbit satellites outside the cluster. The medium-orbit satellites in the cluster are also called cluster heads, and the low-orbit satellites in the cluster are called cluster nodes. These cluster nodes are always connected or idle with the cluster head in the cluster, and can be provided with service support by the cluster head at any time.

[0067] However, due to the different orbital heights of satellites in different orbits within the cluster, their orbital periods around the earth are not exactly the same. Therefore, there must be a long period of time after a certain moment when the medium-orbit satellite and the low-orbit satellite in the cluster are not visible to each other. In this case, power resources are still allocated to the low-orbit satellite, which not only increases the ineffective power consumption of the low-orbit satellite itself, but also causes serious interference to the low-orbit satellites in other clusters. Therefore, the low-orbit satellite in the blind spot of the medium-orbit satellite is placed outside the cluster to let it find a new cluster to join. At the same time, the original cluster head no longer provides power resource allocation to the low-orbit satellite, which is very critical to improving the overall performance of the network.

[0068] When the low-orbit satellites are in the blind spot of each other, the low-orbit satellite is placed outside the cluster, that is, the low-orbit satellite is transferred from the connected state (or idle state) to the switching state. The failure to transfer the state in advance will cause the real-time service transmission to be interrupted and the delay to increase. Conversely, if the low-orbit satellite is placed outside the cluster in advance, it may not be able to find a suitable new cluster to join, so that the satellite is temporarily placed outside any cluster, that is, in a dormant state, and the remaining capacity is not fully utilized, resulting in a waste of satellite resources. Therefore, for various dynamic scenarios of intersatellite networks, a reasonable power allocation strategy is constructed so that it can be applied to complex and realistic satellite space network dynamic topology scenarios. This is of great significance to improving the robustness of satellite networks and ensuring their good performance.

[0069] Therefore, the present invention mainly solves the following technical problems:

[0070] 1. For the same low- and medium-orbit satellite network, mathematical modeling is conducted on the power resource allocation problem with high dynamic characteristics between satellites. By designing a reasonable penalty function, different utility penalties are imposed on low-orbit satellites in different states within the cluster, and low penalties are imposed on low-orbit satellites with good channel conditions within the cluster, while particularly large penalties are imposed on low-orbit satellites that are about to be unable to communicate within the cluster. The purpose is to adaptively match low-orbit satellites in complex heterogeneous states through the above penalty mechanism and to reasonably allocate power resources to them.

[0071] 2. Under the condition of limited satellite resources, when should the operating low-orbit satellite leave the cluster to effectively solve the problem of maximizing the residual value of the low-orbit satellite; at the same time, it can avoid the waste of power resources caused by the low-orbit satellite in the cluster and the interference with other low-orbit satellites in the cluster.

[0072] like Figure 2 The figure shows the inter-satellite operation model of the same medium and low orbit satellite network. Figure 2 As shown in the figure, in the same medium and low orbit satellite network, let the low orbit satellite LEO 1 ,LEO 2 , …, LEO n As participants, the medium-orbit satellite α (medium-orbit satellite is also denoted as MEO) is the manager, and the orbital radius of the medium-orbit satellite is r α , the orbital radii of the low-orbit satellites are r 1 、r 2 ,…,r n , the geocentric angles between the medium-orbit satellite and the low-orbit satellite are Although the orbits of medium and low orbit satellites are not strictly circular, since the inter-satellite link distance can reach several thousand kilometers, this small error does not affect the applicability of the model. i The square of the intersatellite link distance to the medium-orbit satellite α satisfy:

[0073]

[0074] Among them, r α is the orbital radius of the medium-orbit satellite, r i is the orbital radius of the i-th LEO satellite in the cluster, t k refers to the sampling time at the current time slot k, k∈(1, 2, 3, ...), is the geocentric angle between the medium-orbit satellite and the i-th low-orbit satellite in the cluster at the current time slot k.

[0075] The medium-orbit satellite provides service support to n low-orbit satellites in the network at the same time, and the effective transmission power of the n low-orbit satellites is received in the current time slot k.

[0076] Among them, the effective transmission power p of the i-th low-orbit satellite received by the medium-orbit satellite in the current time slot k is ri (t k )for:

[0077]

[0078] Among them, p ti (t k ) is the transmission power of the i-th low-orbit satellite in the cluster at the current time slot k, G ti is the transmission power gain of the i-th LEO satellite in the cluster, G r is the gain of the medium-orbit satellite receiving the power sent by the low-orbit satellite, λ is the wavelength of the transmitted signal, It is the square of the inter-satellite link distance between the i-th low-orbit satellite LEOi and the medium-orbit satellite α in the cluster.

[0079] Because the coefficient Only the sampling time t on the current time slot k k Therefore, when the time slot is determined, the effective transmission power p ri (t k ) is a constant. At this time, the effective transmission power p of the i-th low-orbit satellite received by the medium-orbit satellite in the current time slot k is ri (t k ) can be simplified to:

[0080] p ri (t k )=h i (t k ) ti (t k ), 1≤i≤n,

[0081]

[0082] Let the inter-satellite link gain parameter h(t k )=[h 1 (t k ) 2 (t k )…h n (t k )],h i (t k ) represents the i-th LEO satellite in the cluster i Intersatellite link gain parameters between satellites and medium-orbit satellites.

[0083] Since the orbits of medium and low orbit satellites are fixed, at any given moment, the link distance between medium and low orbit satellites can be calculated by the above The calculation formula is determined, therefore, the link distance set between medium and low orbit satellites in different dynamic scenarios can be obtained by statistics. The link distance set between medium and low orbit satellites in different dynamic scenarios is shown in Table 1.

[0084] Table 1 Link distances between medium and low orbit satellites in different dynamic scenarios

[0085]

[0086] It is known that the operation period of low-orbit satellites is shorter than that of medium-orbit satellites, and the corresponding angular velocity of low-orbit satellites is greater than that of medium-orbit satellites. It is known that the intersatellite link gain parameters of medium-orbit and low-orbit satellites corresponding to different dynamic scenes in the same time slot are different. The complete trajectory of a single low-orbit satellite relative to a medium-orbit satellite includes sub-process 1 from entering the visible area of ​​the medium-orbit satellite to approaching the medium-orbit satellite, then to sub-process 2 away from the medium-orbit satellite, and finally to sub-process 3 entering the blind spot of the medium-orbit satellite. The operation trajectory of a low-orbit satellite that can communicate normally should include sub-process 1 or sub-process 2. Therefore, sub-process 1 and sub-process 2 constitute a set of communicable trajectories of medium-orbit and low-orbit satellites, and any or all elements of the set can be used as a communicable operation scenario of a single low-orbit satellite. Then, several low-orbit satellites in the cluster are combined to form a complex heterogeneous scenario. When the low-orbit satellite enters the visible area of ​​the medium-orbit satellite, there will inevitably be a period of time when the link distance between the medium and low-orbit satellites gradually approaches. This is sub-process 1. Then at a certain moment, the link distance between the medium and low-orbit satellites reaches the shortest, and then the link distance between the medium and low-orbit satellites will gradually increase. This is sub-process 2.

[0087] As shown in Table 1, in the first dynamic scenario, the distance between the low-orbit satellite and the medium-orbit satellite changes from "far" to "near" and then to "far". "Far" to "near" means that the low-orbit satellite starts to approach the medium-orbit satellite from a position far away from the medium-orbit satellite, corresponding to sub-process 1. The subsequent "near" to "far" means that the low-orbit satellite starts to move away from the medium-orbit satellite from a position close to the medium-orbit satellite, corresponding to sub-process 2.

[0088] In the second or third dynamic scenario, the distance between the low-orbit satellite and the medium-orbit satellite changes from "far" to "near", indicating that the low-orbit satellite starts to approach the medium-orbit satellite from a position far away from the medium-orbit satellite, corresponding to sub-process 1; and from "near" to "far", indicating that the low-orbit satellite starts to move away from the medium-orbit satellite from a position close to the medium-orbit satellite, corresponding to sub-process 2.

[0089] As long as in a certain time slot, the link distance between each low-orbit satellite and the medium-orbit satellite is the same as the order of "near" and "far" of each low-orbit satellite from the medium-orbit satellite MEO in a row of the figure, it is regarded as the same dynamic scene as that row.

[0090] For example, if the intersatellite link distance of low-orbit satellite 1 relative to medium-orbit satellites changes from far to near and then to far during the observation period, and the intersatellite link distance of low-orbit satellite 2 relative to medium-orbit satellites also changes from far to near and then to far, etc., then it can be regarded as the first dynamic scene. Similarly, if the distance of low-orbit satellite 1 relative to medium-orbit satellites changes from near to far during the observation period, and the distance of low-orbit satellite n relative to medium-orbit satellites changes from far to near, then it belongs to the second dynamic scene.

[0091] Each dynamic scenario requires that all low-orbit satellites are in the visible area. The clustering removal scheme proposed in this patent will place low-orbit satellites with insufficient remaining visibility time and utility below the threshold out of the cluster in advance, and will not wait until the low-orbit satellites have entered the non-visible area before removing them from the cluster.

[0092] It should be noted that for the same dynamic scenario, the operation of different satellites is not necessarily completely synchronized. For example, for the first dynamic scenario, low-orbit satellite 1 may move from far to near. When it has reached the closest point and then starts to move away, low-orbit satellite 2 may not have yet reached the closest point. Whether it can be completely synchronized depends on the orbits of the satellites. The inter-satellite link power allocation method for heterogeneous scenarios of the present invention is applicable to both synchronous and asynchronous situations between satellites.

[0093] The link distance sets obtained for different dynamic scenes can be used to determine the remaining visible time of the initial low-orbit satellite in steps 1 and 12 below. In other words, different dynamic scenes correspond to different remaining visible times of the initial low-orbit satellite.

[0094] Therefore, the signal-to-noise ratio of n low-orbit satellites in the cluster at the current time slot k is defined as:

[0095]

[0096] Among them, h i (t k ) is the i-th LEO satellite in the cluster iThe intersatellite link gain parameter between the satellite and the medium-orbit satellite, p ti (t k ) is the transmission power of the i-th LEO satellite in the cluster at the current time slot k, σ 2 It is the white noise power in the satellite signal transmission space. That is, increasing the transmission power of a low-orbit satellite will increase its own signal-to-noise ratio, but will cause interference to other low-orbit satellites and reduce their signal-to-noise ratios.

[0097] The normalized throughput C(t k ) is defined as:

[0098] C(t k )=[C 1 (t k ) C 2 (t k )…C n (t k )]

[0099] =β·[log(1+ξ 1 (t k )) log(1+ξ 2 (t k ))…log(1+ξ n (t k ))]

[0100] Among them, C i (t k ) is the normalized throughput of the i-th LEO satellite in the cluster at the current time slot k, β is the normalization factor, ξ i (t k ) is the signal-to-noise ratio of the i-th LEO satellite in the cluster at the current time slot k, ξ i (t k )=p ti (t k ) / σ 2 , p ti (t k ) is the transmission power of the i-th LEO satellite in the cluster at time slot k, σ 2 is the white noise power in the satellite signal transmission space.

[0101] The normalization factor β is:

[0102] β=1 / log(1+ξ max ),

[0103] Among them, ξ max =p max / σ 2 , p max It is the maximum transmission power of a low-orbit satellite.

[0104] The normalized energy consumption E(t k ) is defined as:

[0105]

[0106] Δt is the time slot interval, p ti (t k ) is the transmission power of the i-th LEO satellite in the cluster at the current time slot k, and γ is the normalization factor.

[0107] The normalization factor γ is:

[0108]

[0109] Among them, p max is the maximum transmission power of the low-orbit satellite, t max It is the longest time a low-orbit satellite can operate at maximum transmit power.

[0110] The remaining visible time R(k) of n LEO satellites in the cluster to the ME satellite in the current time slot k is defined as

[0111] R(k)=[R 1 (t k ) R 2 (t k )…R n (t k )]

[0112] =[t 10 -(k-1)Δt t 20 -(k-1)Δt…t i0 -(k-1)Δt…t n0 -(k-1)Δt]

[0113] Among them, t i0 is the initial remaining visible time of the i-th low-orbit satellite to the medium-orbit satellite in the cluster. The sampling time in the first time slot is recorded as the initial remaining visible time. In the current time slot, the k-th low-orbit satellite has experienced (k-1)Δt.

[0114] The initial remaining visible time of the i-th low-orbit satellite to the medium-orbit satellite in the cluster is determined by the dynamic scene in the first observation time slot and the inter-satellite link distance between the i-th low-orbit satellite and the medium-orbit satellite. In the first time slot (i.e., the first observation time slot), the low-orbit satellite may have moved to a certain position in the middle, such as starting to move away from the medium-orbit satellite. Therefore, the present invention realizes the determination of the initial remaining visible time of the low-orbit satellite to the medium-orbit satellite by distinguishing different dynamic scenes.

[0115] According to the two-body motion cycle and spherical geometry knowledge in celestial motion, the initial remaining visible time t of the i-th low-orbit satellite in the cluster to the medium-orbit satellite is i0 The value range of is determined by the intersatellite link distance between the ith LEO satellite and the medium-orbit satellite. The initial remaining visible time t of the ith LEO satellite in the cluster to the medium-orbit satellite is i0 The value range of is:

[0116] 0≤t i0 ≤|t i2 -t i1 |,

[0117] Among them, t i1 ,t i2 is the time when the visibility of the i-th low-orbit satellite switches between two consecutive times, t i1 represents the time when the i-th low-orbit satellite enters the visible area of ​​the medium-orbit satellite, t i2 It indicates the time when the i-th low-orbit satellite leaves the visible area of ​​the medium-orbit satellite.

[0118] The above formula gives the initial remaining visibility time t of the i-th low-orbit satellite to the medium-orbit satellite in the cluster: i0 The specific value position of the remaining visible time in the value range depends on the first time slot to be observed, and is not always calculated from the starting point. Therefore, the specific value position of the remaining visible time in the value range corresponds to the sufficient or insufficient value in the dynamic scene in Table 2.

[0119] Among them, |t i2 -t i1 | is the maximum remaining visible time, that is, the time difference between two consecutive visibility switching moments of the i-th low-orbit satellite. Considering that all low-orbit satellites are about to enter the visible area or the invisible area, their inter-satellite link distance is certain, so we can determine |t i2 -t i1 |.

[0120] Among them, the inter-satellite link distances of the medium and low orbit satellites at the two visibility switching moments all meet the following requirements:

[0121]

[0122] r α 、r i are the orbital radii of the medium-orbit satellite and the i-th low-orbit satellite, and r is the radius of the earth.

[0123] When the visible line between medium and low-orbit satellites is tangent to the earth's spherical surface, it indicates that the medium and low-orbit satellites are about to enter the visible area, or are about to enter the invisible area.

[0124] Corresponding to the scenarios shown in Table 1, the remaining visible time sets of different operating scenarios of low-orbit satellites in the cluster are shown in Table 2. As time goes by, the remaining visible time of low-orbit satellites in the cluster gradually decreases.

[0125] Table 2 Dynamic scenarios of remaining visible time of low-orbit satellites

[0126]

[0127] As shown in Table 2, when the low-orbit satellite is still approaching the medium-orbit satellite, it means that the remaining visible time is sufficient, because the low-orbit satellite has a period of moving away from the medium-orbit satellite before finally entering a state of being invisible to each other.

[0128] The insufficiency corresponds to the process of the low-orbit satellite moving away from the medium-orbit satellite, so the remaining visible time is insufficient compared to the former.

[0129] Because L αi (t k ) is a convex function of time, and has a minimum extreme point. Let dL αi (t) / dt=0, the obtained t is t min , t min Refers to the moment when the inter-satellite link distance between medium and low-orbit satellites is the shortest.

[0130] Whether the remaining time is sufficient or not is defined as:

[0131]

[0132] Among them, t min Refers to the time when the inter-satellite link distance between medium and low orbit satellites is the shortest, t i1 is the time when the i-th low-orbit satellite enters the visible area of ​​the medium-orbit satellite, t i2 is the moment when the i-th low-orbit satellite leaves the visible area of ​​the medium-orbit satellite.

[0133] The principle of the intersatellite link power allocation method for heterogeneous scenarios of the present invention (i.e., the power allocation principle within the visible range of medium and low orbit satellites) is to maximize the overall utility of the same medium and low orbit satellite networking network. The utility function is composed of a benefit function and a cost function, wherein the benefit size is related to the throughput, and the cost size is related to the energy consumption and the remaining visible time. From the throughput definition formula, it is known that the improvement of the signal-to-noise ratio of low-orbit satellites will increase their own throughput, thereby increasing the benefits. From the energy consumption definition formula, it is known that the low-orbit satellite will increase its own energy consumption with a larger transmission power, thereby increasing the cost.

[0134] The remaining visible time penalty function Cost of n low-orbit satellites in the cluster at the current time slot k time (t k ) is defined as:

[0135]

[0136] Among them, the coefficient k i1 is the remaining visible time convergence coefficient, i∈(1, 2, …, n), which determines the speed of convergence of the remaining visible time cost penalty function, k i2 Defined as the remaining visible time penalty coefficient, k is the ordinal number of the current time slot, and Δt is the time slot interval.

[0137] Remaining visible time convergence coefficient k i1 The value of depends on the orbital period of the satellite and satisfies:

[0138] k i1 ∝1 / T i (k i1 >0)

[0139] Ti is the orbital period of the i-th low-orbit satellite in the cluster. That is, the smaller the period, the faster the corresponding cost function converges, implying that the stronger the spatial time-varying characteristics of the low-orbit satellite, the greater the adjustment range of its strategy in different time slots.

[0140] The orbital period of the K low-orbit satellites in the cluster at the current time slot k is determined according to the orbital radius of the low-orbit satellite. Therefore, the corresponding period T is selected when selecting the low-orbit satellite members in the cluster. i It has been determined that:

[0141]

[0142] Among them, r i is the orbital radius of the ith low-orbit satellite, G is the gravitational constant, and M is the mass of the Earth.

[0143] Remaining visible time penalty coefficient k i2 , i∈(1, 2, ..., n), determines the size of the penalty, the remaining visible time penalty coefficient k i2 The value of depends on the current satellite throughput and satisfies:

[0144] k i2 ∝1 / log(1+ξ i (t k ))(k i2 >0)

[0145] Among them, log(1+ξ i (t k )) is the throughput of n low-orbit satellites in the cluster at the current time slot k, ξ i (t k) is the signal-to-noise ratio of the ith low-orbit satellite in the cluster at the current time slot k, that is, the larger the throughput, the smaller the corresponding penalty of the cost function, implying that the better the quality of the low-orbit satellite channel, the less inclined to impose a large time cost penalty on it, that is, to give full play to the margin of the low-orbit satellite in the remaining time.

[0146] When the remaining visible time of the i-th low-orbit satellite is zero, its remaining visible time penalty function Cost time (t k ) satisfies the absolute value of:

[0147]

[0148] That is, the cost penalty function is subject to a particularly large utility penalty based on the remaining visible time.

[0149] According to the above normalized energy consumption E(t k ) Define the energy cost penalty function Cost on the current time slot k energy (t k ).

[0150] Energy consumption penalty function Cost at the current time slot k energy (t k )for:

[0151] Cost energy (t k )=[ηE 1 (t k ) ηE 2 (t k )…ηE n (t k )],

[0152] Among them, η is the energy consumption penalty factor, E i (t k ) is the normalized energy consumption of the i-th low-orbit satellite in the current time slot k.

[0153] because In order to increase the impact of energy consumption penalty on utility, η is taken as a larger value; but at the same time, considering that the low-orbit satellite is in a scenario with good channel conditions and low interference to other users, the satellite is more inclined to transmit at maximum power to increase throughput, so To get the energy consumption penalty factor η. K 能耗 It is an empirical value, usually 0.5 (i.e. K 能耗≈0.5). If the value is greater than 0.5, the energy consumption cost caused by the low-orbit satellite when transmitting at maximum power will also be greater. In order to ensure the overall utility, the satellite will not tend to transmit at maximum power, but the satellite's capabilities when channel conditions are good will not be fully exploited. On the contrary, if the value is smaller than 0.5, the energy consumption cost will be reduced. The low-orbit satellite is particularly inclined to transmit at maximum power, but it will cause two negative effects. First, always transmitting at the maximum transmit power will increase energy consumption. Second, it will cause great interference to other low-orbit satellites, so the overall utility will not be very good.

[0154] The normalized throughput C(t k ) defines the profit function on the current time slot k. The profit function R on the current time slot k throughput (t k )for:

[0155] R throughput (t k )=[λC 1 (t k ) λC 2 (t k )…λC n (t k )],

[0156] Among them, λ is the throughput benefit factor. Throughput is the only reward factor for utility. Here, λ=1.

[0157] Then the utility value u(t k )for:

[0158] u(t k )=[u 1 (t k ) 2 (t k )…u n (t k )]

[0159]

[0160] Among them, u i (t k ) is the utility value of the i-th low-orbit satellite in the cluster at the current time slot k, R throughput (t k ) is the profit function at the current time slot k, Cost energy (t k ) is the energy cost penalty function at the current time slot k, Cost time (t k ) is the remaining visible time cost penalty function.

[0161] In order to maximize the overall utility of the medium and low orbit satellite network, that is, to make the overall utility value of n low orbit satellites in the cluster at the current time slot k The maximum power value corresponding to the best utility of the low-orbit satellite in the current time slot k needs to satisfy:

[0162]

[0163]

[0164]

[0165]

[0166] in, The meaning of is the optimal strategy of n low-orbit satellite participants in the current time slot k, that is, the optimal transmission power. refers to the optimal transmission power of the i-th low-orbit satellite in the current time slot k, refers to the optimal transmission power of the other n-1 low-orbit satellites in the cluster except the i-th low-orbit satellite at the current time slot k, p ti (t k ) refers to the transmission power of the i-th low-orbit satellite in the current time slot k, It refers to the transmission power of any LEO satellite in the cluster except the i-th LEO satellite in the current time slot k. Refers to the transmission power of low-orbit satellites. u(t k ) is the utility value of n low-orbit satellites in the cluster at the current time slot k, which is in the form of a queue, representing the utility of each member in turn; It refers to the overall utility of the ith low-orbit satellite and all low-orbit satellites except the ith low-orbit satellite in the current time slot k.

[0167] formula It is used to ensure that an optimal transmission strategy is achieved. The formula means that when the game participants determine the optimal strategy, no player can improve the overall utility by changing their own strategy (transmission power). It is used to remove participants whose utility is lower than the de-clustering threshold. Satisfying the above four formulas can maximize the overall utility value, so the algorithm core of the present invention is to minimize the distance between the current strategy and the optimal strategy.

[0168] The value of the clustering threshold of the i-th low-orbit satellite in the cluster depends on the weight of the satellite throughput benefit and energy consumption penalty at the current time slot k, satisfying:

[0169]

[0170] Among them, μ is a weight factor, which is used to balance the impact of throughput benefits and energy consumption penalties. If the throughput weight is larger, the clustering strategy is more conservative, and the low-orbit satellite is less likely to be clustered. On the contrary, if the energy consumption weight is larger, the clustering strategy is more aggressive, and the low-orbit satellite is more likely to be clustered. The value range of μ is 0.01-0.5.

[0171] like Figure 3 FIG. 2 is a flow chart of the inter-satellite link power allocation method for heterogeneous scenarios of the present invention. Figure 3 As shown, based on the above principle, the inter-satellite link power allocation method for heterogeneous scenarios of the present invention includes:

[0172] Step S1: Set the dynamic scene of low-orbit satellites relative to medium-orbit satellites, initialize the number of low-orbit satellites K in the current cluster, and set its initial value to N by default, so that K = N; set the current time slot k = 1 (that is, set the initial time slot), set the time slot interval Δt; set the initial remaining visible time t of the N low-orbit satellites in the cluster to the medium-orbit satellite, and set t = [t 10 t 20 …t N0 ], t i0 is the initial remaining visible time of the i-th low-orbit satellite to the medium-orbit satellite in the cluster; set the transmission power range of the low-orbit satellite [p min , p max ], set the transmission power step length Δp, and obtain the number of power levels M = (p max -p min ) / Δp, and obtain the transmission power vector of the low-orbit satellite: [p min p min +Δp…p max -Δp p max ].

[0173] Among them, parameter K is a variable, which refers to the number of low-orbit satellites in the cluster, and N is a constant, which refers to the initial value assigned to parameter K.

[0174] As mentioned above, the initial remaining visible time t of the i-th LEO satellite to the ME satellite in the cluster is i0 The value range of t depends on the relative positions of the low-orbit satellites at the initial moment, so t i0 The value range is from 0 to |t i2 -t i1 |Any value. If it is 0, it means that the low-orbit satellite has just entered the invisible area and the remaining visible time is 0. If it is |t i2 -t i1 | indicates that the low-orbit satellite has just entered the visible area and the remaining visible time is the maximum. It ranges from 0 to |t i2 -t i1|Any value in the middle means that the low-orbit satellite is at any position between entering the visible area and leaving the visible area.

[0175] The initial remaining visible time t of the i-th low-orbit satellite to the medium-orbit satellite in the cluster i0 The value of satisfies the following formula:

[0176]

[0177] Among them, t i1 is the time when the i-th low-orbit satellite enters the visible area of ​​the medium-orbit satellite, t i2 is the time when the i-th low-orbit satellite leaves the visible area of ​​the medium-orbit satellite, t min Refers to the time when the inter-satellite link distance between medium and low orbit satellites is the shortest. min It is obtained by the following method: Let dL αi (t) / dt=0, the obtained t is t min .

[0178] Step S2: If K≠0, update the inter-satellite link gain parameter h(t k )=[h 1 (t k ) 2 (t k )…h K (t k )]; According to the orbital period of the K low-orbit satellites in the cluster, the normalized throughput C(t k ) respectively update the remaining visible time convergence coefficient k i1 and the remaining visible time penalty coefficient k i2 ; If K=0, go to step S5.

[0179] Among them, the inter-satellite link gain parameter h(t k )=[h 1 (t k ) 2 (t k )…h K (t k )] is calculated by the following formula:

[0180]

[0181] h i (t k ) represents the inter-satellite link gain parameter between the i-th LEO satellite and the ME satellite in the cluster.

[0182] Step S3: If K=1, go to step S4; otherwise, go to step S6.

[0183] Step S4: output the best utility of the only low-orbit satellite in the current time slot k and its corresponding power value, and proceed to step S11;

[0184] Since there is only one satellite left in the cluster at this time, the interference to other satellites in the cluster is not considered. Therefore, the power value corresponding to the best utility of the low-orbit satellite in the current time slot k is the maximum transmission power of the low-orbit satellite. The best utility of the low-orbit satellite in the current time slot k is based on the maximum transmission power of the low-orbit satellite and the utility value u of the i-th low-orbit satellite in the cluster at the current time slot k. i (t k ) is calculated using the calculation formula.

[0185] That is to say, the utility value u of the i-th low-orbit satellite in the cluster at the current time slot k is i (t k ) Based on the profit function R at the current time slot k throughput (t k ), the energy cost penalty function Cost at the current time slot k energy (t k ) and the remaining visible time penalty function Cost time (t k ) to calculate.

[0186] Step S5: Set the best utility at the current time slot k to zero, and proceed to step S11;

[0187] Step S6: Fix the transmission power p of the i-th low-orbit satellite in the current time slot k ti (t k ), traverse the transmission power of the other K-1 low-orbit satellites in the cluster except the i-th low-orbit satellite The overall utility of the i-th low-orbit satellite and all low-orbit satellites except the i-th low-orbit satellite in the current time slot k is calculated under different combinations of transmission power. Select the maximum value to form the initial overall maximum utility array u of the i-th low-orbit satellite in the current time slot k ini_max_i , the transmission power of each of the K LEO satellites in the corresponding cluster is stored in the initial optimal transmission power response array p of the i-th LEO satellite ini_best_i_q_matrix , the corresponding transmission power of the i-th low-orbit satellite constitutes the initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i ;

[0188] The initial overall maximum utility array u of the i-th low-orbit satellite at the current time slot k is ini_max_i for:

[0189]

[0190] Among them, u in i _max_i_q It refers to the initial overall maximum utility value of the i-th low-orbit satellite under the q-th transmission power combination, 1≤q≤M K-1 , M K-1 represents the number of combinations of the transmission power of the other K-1 low-orbit satellites in the cluster (a total of M K-1 different combinations of transmit powers), M is the number of power levels.

[0191] The initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i for:

[0192]

[0193] Among them, p ini_best_i_q is the initial optimal transmission power of the i-th low-orbit satellite under the q-th transmission power combination.

[0194] The initial overall maximum utility value u of the i-th low-orbit satellite under the q-th transmission power combination ini_max_i_q and the corresponding initial optimal transmission power array p ini_best_i The calculation process is as follows:

[0195] Take a set of transmission power combinations of K-1 low-orbit satellites in the cluster except the i-th low-orbit satellite as the q-th transmission power combination, traverse the different transmission powers of the i-th low-orbit satellite under the condition of the q-th transmission power combination, and calculate the utility value u of the i-th low-orbit satellite in the cluster at the current time slot k i (t k ) and the calculation formula for the overall utility of the ith low-orbit satellite and all low-orbit satellites except the ith low-orbit satellite in the current time slot k Calculate the overall utility of the i-th low-orbit satellite at different transmission powers And select the overall utility The maximum value of is taken as the initial overall maximum utility value u of the i-th low-orbit satellite under the condition of the n-th transmission power combination ini_max_i_q (1≤q≤M K-1 ), the corresponding transmission power of the i-th low-orbit satellite is the initial optimal transmission power p of the i-th low-orbit satellite under the q-th transmission power combination ini_best_i_q At the same time, the transmission power of each of the K low-orbit satellites in this conditional cluster is stored in the initial optimal transmission power response array p of the i-th low-orbit satellite ini_best_i_q_matrix In the example, the initial optimal transmission power p for the low-orbit satellite i under the qth transmission power combination is ini_best_i_n Supplement.

[0196] By analogy, M K-1 After traversing the combinations of different transmission powers, the initial overall maximum utility array u of the i-th low-orbit satellite in the current time slot k is obtained. ini_max_i , and the initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i .

[0197] Step S7: i=i+1, and return to step S6 until all low-orbit satellites in the cluster are traversed, and the initial overall maximum utility set u is obtained. ini_max and the corresponding initial optimal transmit power response set, and continue to execute step S8;

[0198] After traversing all K low-orbit satellites in the cluster, we can obtain the initial overall maximum utility set of K low-orbit satellites, denoted as u ini_max =[u ini_max_1 u ini_max_2 …u ini_max_K ] T , which is a K-row M K-1 A matrix of columns, where the i-th row of M K-1 The element represents the M position of the i-th low-orbit satellite among the other K-1 satellites. K-1 The initial overall maximum utility value u corresponding to the different transmission power combinations ini_max_i_q The corresponding initial optimal transmission power response set includes the M of each low-orbit satellite. K-1 The initial optimal transmission power response array p corresponding to the different transmission power combinations ini_best_i_q_matrix .

[0199] Step S8: Calculate the initial optimal transmit power response array of one of the low-orbit satellites in the initial optimal transmit power response set as the reference satellite (such as the p of the first low-orbit satellite). ini_best_1_m_matrix ) and the initial optimal transmission power response array of any other low-orbit satellite (such as p ini_best_i_n_matrix ), the transmission power corresponding to the minimum distance between the matrices of the reference satellite and the i-th low-orbit satellite is selected as the power output result of the i-th low-orbit satellite after the game, so as to obtain the power output result of each low-orbit satellite after the game, that is, the Nash equilibrium solution.

[0200] The step S8 comprises: taking the first low-orbit satellite as a reference satellite, selecting an initial optimal transmission power response array p corresponding to different transmission power combinations of the reference satellite and the i-th low-orbit satellite ini_best_i_q_matrix The initial optimal transmission power response array pair p with the smallest distance between ini_best_1_m_matrix 、p ini_best_i_n_matrix, and output the initial overall maximum utility values ​​corresponding to the transmission power combinations m and n of the initial optimal transmission power response array pair, respectively, as the power output results p of the reference satellite after the game ini_best_1 (m) and the power output result p after the game with the i-th low-orbit satellite ini_best_i (n), thereby obtaining the initial overall maximum utility of the ith low-orbit satellite and the corresponding initial optimal transmission power, and realizing one round of game; then, the same logic is applied under the condition of changing the size of i to calculate the power output results of other low-orbit satellites in the cluster and the reference satellite after the game.

[0201] Among them, the initial optimal transmission power response array p corresponding to the combination of different transmission powers of the reference satellite and the i-th low-orbit satellite is selected by the following formula: ini_best_i_q_matrix The initial optimal transmit power response array pair with the smallest distance between them is:

[0202] min||p ini_best_1_m_matrix -p ini_best_i_n_matrix ||, 1≤m, n≤M K-1 .

[0203] In each round of game, the power output result p of the reference satellite after the game is ini_best_1 (m) should be the same.

[0204] Therefore, steps S6-S8 obtain the power output results of each low-orbit satellite after the game based on the overall utility game.

[0205] Step S9: Calculate the utility value and cluster-leaving threshold of each low-orbit satellite in the cluster at the current time slot k using the power output results of each low-orbit satellite in the cluster after the game. If the utility value of any low-orbit satellite is lower than the cluster-leaving threshold, the low-orbit satellite will be de-clustered. At this time, K=Kn, where n refers to the number of satellites whose utility value is lower than the cluster-leaving threshold.

[0206] As mentioned above, the utility value u of the i-th low-orbit satellite in the cluster at the current time slot k is i (t k )for:

[0207]

[0208] Among them, R throughput (t k ) is the profit function at the current time slot k, Cost energy (t k ) is the energy cost penalty function at the current time slot k, Cost time (t k ) is the remaining visible time cost penalty function.

[0209] The clustering threshold of the i-th low-orbit satellite in the cluster is:

[0210]

[0211] Among them, μ is a weight factor, which is used to balance the impact of throughput benefit and energy consumption penalty. If the throughput weight is larger, the declustering strategy is more conservative, and the low-orbit satellite is less likely to decluster. Conversely, if the energy consumption weight is larger, the declustering strategy is more aggressive, and the low-orbit satellite is more likely to decluster.

[0212] The value range of μ is 0.01-0.5.

[0213] Step S10: The power output results of all low-orbit satellites in the cluster that have not left the cluster after the game are taken as their optimal transmission power, and the sum of the utility values ​​of all low-orbit satellites in the cluster at the current time slot k is calculated and taken as the optimal overall utility of the current time slot.

[0214] Step S11: Enter the next time slot and set k=k+1; then, if K>0, update the remaining visible time of all low-orbit satellites in the cluster (i.e., t=t-Δt), and then enter step S12: If K=0, that is, there is no satellite in the cluster, then there is no need to update the remaining visible time of the current low-orbit satellite in the cluster, and directly enter step S12;

[0215] Step S12: Initialize the remaining visible time of all low-orbit satellites newly added to the cluster, and use K+1 as the updated K, where 1 is the number of low-orbit satellites newly added to the cluster in the new current time slot.

[0216] It should be noted that, in the present invention, the l new low-orbit satellites in the cluster may not necessarily be added as soon as they enter the visible area, but may be added in the middle. i0 The setting is still an interval, from 0 to |t i2 -t i1 |, the specific setting depends on the position of the low-orbit satellite.

[0217] The feature of the present invention is that it is not limited to whether the satellite is added as soon as it enters the visible area, but is applicable to a variety of different scenarios. In other words, this patent is applicable to the satellite joining from any visible area. For details, please refer to the complex and heterogeneous dynamic scenes in Table 2.

[0218] Step S13: If all time slots have been traversed, the process ends, otherwise it returns to step S2.

[0219] The intersatellite link power allocation method for heterogeneous scenarios of the present invention combines the low-orbit satellite's own orbit and throughput to design the remaining visible time cost penalty function Cost time (t k), a greater utility penalty is imposed on the LOR satellite with less remaining visible time. If the channel quality of the LOR satellite is relatively good, the utility penalty is reduced accordingly. If the LOR satellite is in the blind spot of the medium-orbit satellite, a particularly large utility penalty is imposed, thereby solving the problem of power resource allocation for LOR satellites in a complex heterogeneous state.

[0220] In addition, the intersatellite link power allocation method for heterogeneous scenarios of the present invention designs the clustering threshold of the utility of low-orbit satellites on the basis of the game of overall utility. The size of the clustering threshold depends on the weight of the current satellite throughput benefit and the energy consumption penalty, so that the low-orbit satellite participants with utility lower than the threshold are placed in the clustering state to ensure the overall utility of the low-orbit satellite network in the cluster. If the throughput weight is larger, the clustering strategy is more conservative, and the low-orbit satellite is less likely to be clustered. On the contrary, if the energy consumption weight is larger, the clustering strategy is more radical, and the low-orbit satellite is more likely to be clustered.

[0221] In summary, the inter-satellite link power allocation method for heterogeneous scenarios of the present invention, by designing a reasonable utility reward and punishment function (mainly including designing a satellite remaining visible time cost penalty function) and an effective utility out-of-cluster threshold, enables the medium and low-orbit satellite networking network to maximize the performance and online rate of the low-orbit satellites in the cluster, while avoiding the abnormal situation that the low-orbit satellites in the cluster are in a communication blind spot but still remain in the cluster, thereby optimizing the overall performance of the medium and low-orbit satellite network.

[0222] Through theoretical proof, the present invention is based on the game power allocation scheme of the remaining visible time cost penalty function. The cost function comprehensively considers the satellite's own trajectory and throughput, and the low-orbit satellite members in the cluster conduct utility games in a non-cooperative manner. Compared with the traditional greedy power allocation and water injection power allocation schemes, it has better performance improvement in a variety of dynamic scenarios. At the same time, on this basis, a low-orbit satellite utility clustering threshold scheme is designed to ensure a certain satellite online rate in the satellite network and reduce the switching frequency of low-orbit satellites. At the same time, through the effect of the remaining visible time cost penalty, the low-orbit satellite that is severely punished can be placed in a clustering state in time, reducing its utility penalty to the whole, so as to optimize the overall performance of the medium and low-orbit satellite network.

[0223] The above is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. The above embodiments of the present invention can also be modified in various ways. All simple, equivalent changes and modifications made according to the claims and the description of the present invention fall within the scope of protection of the claims of the present invention. The contents not described in detail in the present invention are all conventional technical contents.

Claims

1. A method for allocating intersatellite link power in heterogeneous scenarios. It is characterized in that include: Step S1: Set the dynamic scene of low-orbit satellites relative to medium-orbit satellites, initialize the number of low-orbit satellites K in the current cluster; set the current time slot k=1, set the time slot interval Δt; set the initial remaining visible time of N low-orbit satellites in the cluster to the medium-orbit satellite; set the number of power levels M and the transmission power vector of the low-orbit satellites; Step S2: If K≠0, update the inter-satellite link gain parameter h(t k )=[h 1 (t k ) 2 (t k ) … h K (t k )]; According to the orbital period of the K low-orbit satellites in the cluster at the current time slot k, the normalized throughput C(t k ) respectively update the remaining visible time convergence coefficient k i1 and the remaining visible time penalty coefficient k i2 , and then get the remaining visible time penalty function Cost time (t k ); if K = 0, go to step S5; Step S3: If K=1, proceed to step S4; otherwise, proceed to step S6; Step S4: output the best utility of the only low-orbit satellite in the current time slot k and its corresponding power value, and proceed to step S11; The power value corresponding to the best utility of the low-orbit satellite in the current time slot k makes the overall utility of the i-th low-orbit satellite and all low-orbit satellites except the i-th low-orbit satellite in the current time slot k Maximum, satisfies: in, Refers to the power value corresponding to the best utility of the i-th low-orbit satellite in the current time slot k, It refers to the power value corresponding to the best utility of the other n-1 low-orbit satellites in the cluster except the i-th low-orbit satellite at the current time slot k, p ti (t k ) refers to the transmission power of the i-th low-orbit satellite in the current time slot k, It refers to the transmission power of any LEO satellite in the cluster except the i-th LEO satellite in the current time slot k; It refers to the overall utility of the ith LEO satellite and all LEO satellites except the ith LEO satellite in the current time slot k; Among them, the overall utility of the i-th low-orbit satellite and all low-orbit satellites except the i-th low-orbit satellite in the current time slot k is The calculation formula is: u i (t k ) is the utility value of the i-th low-orbit satellite in the cluster at the current time slot k; The utility value u of the i-th low-orbit satellite in the cluster at the current time slot k is i (t k )for: u i (t k )=R throughput_i (t k )-Cost energy_i (t k )-Cost time_i (t k ), R throughput_i (t k ) is the profit function at the current time slot k, Cost energy_i (t k ) is the energy cost penalty function at the current time slot k, Cost time_i (t k ) is the remaining visible time cost penalty function on the current time slot k; Step S5: Set the best utility at the current time slot k to zero, and proceed to step S11; Step S6: Fix the transmission power p of the i-th low-orbit satellite in the current time slot k ti (t k ), traverse the transmission power of the other K-1 low-orbit satellites in the cluster except the i-th low-orbit satellite The overall utility of the ith low-orbit satellite and all low-orbit satellites except the ith low-orbit satellite in the current time slot k is calculated under different combinations of transmission power, and the maximum value is selected to form the initial overall maximum utility array u of the ith low-orbit satellite in the current time slot k ini_max_i , the transmission power of each of the K LEO satellites in the corresponding cluster is stored in the initial optimal transmission power response array p of the i-th LEO satellite ini_best_i_q_matrix , the corresponding transmission power of the i-th low-orbit satellite constitutes the initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i ; Step S7: i=i+1, and return to step S6 until all low-orbit satellites in the cluster are traversed, and the initial overall maximum utility set u is obtained. ini_max and the corresponding initial optimal transmit power response set, and continue to execute step S8; Step S8: Calculate the minimum distance between the initial best transmit power response array of one of the low-orbit satellites used as the reference satellite in the initial best transmit power response set and the initial best transmit power response array of any other low-orbit satellite, and select the transmit power combination corresponding to the minimum distance as the power output result after the game of each low-orbit satellite in the cluster; Step S9: Calculate the utility value and cluster-leaving threshold of each low-orbit satellite in the cluster at the current time slot k using the power output results of each low-orbit satellite in the cluster after the game. If the utility value of any low-orbit satellite is lower than the cluster-leaving threshold, the low-orbit satellite is de-clustered. At this time, K=Kn, where n refers to the number of satellites whose utility value of low-orbit satellites is lower than the cluster-leaving threshold. Step S10: The power output results of all low-orbit satellites in the cluster that have not left the cluster after the game are taken as their optimal transmission power, and the sum of the utility values ​​of all low-orbit satellites in the cluster at the current time slot k is calculated and taken as the optimal overall utility of the current time slot; Step S11: let k=k+1; then, if K>0, update the remaining visible time of all low-orbit satellites in the cluster respectively, and then enter step S12: if K=0, directly enter step S12; Step S12: Initialize the remaining visible time of all low-orbit satellites newly added to the cluster, and use K+1 as the updated K, where 1 is the number of low-orbit satellites newly added to the cluster in the new current time slot; Step S13: If all time slots have been traversed, the process ends, otherwise it returns to step S2.

2. The inter-satellite link power allocation method for heterogeneous scenarios according to claim 1, It is characterized in that In step S1 and step S12, the initial remaining visible time t is: t=[t 10 t 20 … t N0 ], t i0 is the initial remaining visibility time of the i-th LEO satellite to the ME satellite in the cluster; t i0 The value range of is: 0 ≤ t i0 ≤ |t i2 - t i1 |, Among them, t i1 ,t i2 is the time when the visibility of the i-th low-orbit satellite switches between two consecutive times, t i1 represents the time when the i-th low-orbit satellite enters the visible area of ​​the medium-orbit satellite, t i2 represents the time when the i-th low-orbit satellite leaves the visible area of ​​the medium-orbit satellite; In step S1, the transmission power range of the low-orbit satellite is set [p min ,p max ], set the transmission power step length Δp, and obtain the number of power levels M = (p max -p min ) / Δp, the transmission power vector of the low-orbit satellite is obtained as: [p min p min +Δp… p max -Δp p max ].

3. The inter-satellite link power allocation method for heterogeneous scenarios according to claim 1, It is characterized in that The remaining visible time penalty function Cost of n low-orbit satellites in the cluster at the current time slot k time (t k )for: Among them, the coefficient k i1 is the remaining visible time convergence coefficient, i∈(1,2,…,n), which determines how fast the remaining visible time cost penalty function converges, k i2 Defined as the remaining visible time penalty coefficient, k is the ordinal number of the current time slot, and Δt is the time slot interval; k i1 ∝1 / T i (k i1 >0),k i2 ∝1 / log(1+ξ i (t k ))(k i2 >0), Among them, T i is the orbital period of the i-th LEO satellite in the cluster, log(1+ξ i (t k )) is the throughput of n low-orbit satellites in the cluster at the current time slot k, ξ i (t k ) is the signal-to-noise ratio of the i-th LEO satellite in the cluster at the current time slot k; The clustering threshold of the i-th low-orbit satellite in the cluster is: Among them, μ is the weight factor, and the value range of μ is 0.01-0.

5.

4. The inter-satellite link power allocation method for heterogeneous scenarios according to claim 1, It is characterized in that Energy consumption penalty function Cost at the current time slot k energy (t k )for: Cost energy (t k )=[ηE 1 (t k ) ηE 2 (t k ) … ηE n (t k )], Among them, η is the energy consumption penalty factor, E i (t k ) is the normalized energy consumption of the i-th low-orbit satellite in the current time slot k.

5. The inter-satellite link power allocation method for heterogeneous scenarios according to claim 1, It is characterized in that The profit function R at the current time slot k throughput (t k )for: R throughput (t k )=[λC 1 (t k ) λC 2 (t k ) … λC n (t k )], Where λ is the throughput benefit factor, C i (t k ) is the normalized throughput C(t k ); The normalized throughput C(t k )for: C(t k )=[C 1 (t k ) C 2 (t k ) … C n (t k )]=β·[log(1+ξ 1 (t k )) log(1+ξ 2 (t k )) … log(1+ξ n (t k ))], Among them, C i (t k ) is the normalized throughput of the i-th LEO satellite in the cluster at the current time slot k, β is the normalization factor, ξ i (t k ) is the signal-to-noise ratio of the i-th LEO satellite in the cluster at the current time slot k, ξ i (t k )=p ti (t k ) / σ 2 , p ti (t k ) is the transmission power of the i-th LEO satellite in the cluster at time slot k, σ 2 is the white noise power in the satellite signal transmission space.

6. The inter-satellite link power allocation method for heterogeneous scenarios according to claim 1, It is characterized in that The initial overall maximum utility array u of the i-th low-orbit satellite at the current time slot k is ini_max_i for: in, It refers to the initial overall maximum utility value of the i-th low-orbit satellite under the q-th transmission power combination, 1≤q≤M K-1 , M K-1 represents the number of combinations of the transmission power of the other K-1 low-orbit satellites in the cluster, and M is the number of power levels; The initial optimal transmission power array p of the i-th low-orbit satellite ini_best_i for: Among them, p ini_best_i_q is the initial optimal transmission power of the i-th low-orbit satellite under the q-th transmission power combination.

7. The method for allocating power of inter-satellite links for heterogeneous scenarios according to claim 6, It is characterized in that The initial overall maximum utility value of the i-th low-orbit satellite under the q-th transmission power combination and the corresponding initial optimal transmission power array p ini_best_i The calculation process is as follows: Take a set of transmission power combinations of K-1 low-orbit satellites in the cluster except the i-th low-orbit satellite as the q-th transmission power combination, traverse the different transmission powers of the i-th low-orbit satellite under the condition of the q-th transmission power combination, and calculate the utility value u of the i-th low-orbit satellite in the cluster at the current time slot k i (t k ) and the calculation formula for the overall utility of the ith low-orbit satellite and all low-orbit satellites except the ith low-orbit satellite in the current time slot k Calculate the overall utility of the i-th low-orbit satellite at different transmission powers The maximum value of the overall utility is selected as the initial overall maximum utility value of the i-th low-orbit satellite under the condition of the n-th transmission power combination. At this time, the corresponding transmission power of the i-th low-orbit satellite is the initial optimal transmission power p of the i-th low-orbit satellite under the q-th transmission power combination. ini_best_i_q At the same time, the transmission power of each of the K low-orbit satellites in the cluster is stored in the initial optimal transmission power response array p of the i-th low-orbit satellite ini_best_i_q_matrix In the example, the initial optimal transmission power p for the low-orbit satellite i under the qth transmission power combination is ini_best_i_n Supplement.

8. The method for allocating inter-satellite link power for heterogeneous scenarios according to claim 1, It is characterized in that The step S8 includes: taking the first low-earth orbit satellite as a reference satellite, and selecting the initial optimal transmit power response arrays p corresponding to the combinations of different transmit powers of the reference satellite and the i-th low-earth orbit satellite respectively ini_best_i_q_matrix for the pair of initial optimal transmit power response arrays p with the smallest distance therebetween ini_best_1_m_matrix and p ini_best_i_n_matrix , and outputting the initial overall maximum utility values corresponding to the combinations m and n of the transmit powers of the pair of initial optimal transmit power response arrays respectively as the power output results p ini_best_1 (m) after the game of the reference satellite and the power output result p ini_best_i (n) after the game of the i-th low-earth orbit satellite; subsequently, under the condition of changing the value of i, and so on, to calculate the power output results after the game of other low-earth orbit satellites and the reference satellite within the cluster.

Images